National Conference on Computer Aided Analysis and Design of Structures (CAADS-2016)
Reliability Analysis of Flanged RC Beams
in Limit State of Cracking
1
Dr. K Manjunath, 2Meghana Bharadwaj R.
1
Professor and HOD, Civil Engg Dept.,MCE, Hassan. kmnrpur@gmail.com,
2
PG Student (CADS), MCE, Hassan. mbharadwaj48@gmail.com
Abstract
With a view to characterize randomness in the
surface crack width of RC flanged beam, an attempt has
been made to establish the probability distribution. IS 456-
2000 demonstrates the concept of characteristic value for
material strength and loads. But the same concept has not
been extended to crack width of RC members, even though
it is a well known fact that the crack width is a random
variable.
Since the cracking of a RC members is a highly
random process and hence it is not possible to predict the
crack width without associating such prediction with some
probability of being exceeded. This necessitates the
probability modeling of crack width. The concept of
characteristic crack width can be formulated only after
establishing the probability distribution the crack width
follows. For example if the crack width follows a normal
distribution then the characteristic crack width is given by,
Wch = Wavg + 1.645
Where, is the standard deviation of crack
width.
This characteristic crack width has a probability
of 5% of being exceeded. The statistical parameters of
crack width like mean and standard deviation are required.
In the above equation, a constant 1.645 holds good only for
a normal distribution.
A very large number of samples are needed for
the purpose of computing the statistics of crack width.
However to obtain such a large body of statistics a large
scale experimentation is to be planned at an exorbitant cost.
Digital simulation will solve this problem as it corresponds
to artificial experimentation on a computer.
IS 456-2000 has suggested an equation for
computing the surface crack width. This equation
identifies several basic parameters which influence the
crack width. The statistics of these basic variables are taken
from the literature. Monte Carlo simulation technique is
adopted and sample values of the surface crack width for a
beam of known dimensions are generated. The generated
values are statistically analyzed and probabilistically
modeled. Using the facility on MATLAB platform a
suitable probability distribution is tried to describe the
randomness in the generated data. Both singly reinforced
Department of Civil Engineering, SDMCET, Dharwad-580002
Paper ID: CAD34
and doubly reinforced T-Beams are considered for the
study. Formulations are made for different location of
neutral axis.
The effect of changing each of the basic
parameters is studied. The corresponding statistics,
probability distribution fitting and histograms are
presented. Once the probability distribution is identified,
the characteristic value of the crack width is computed at a
known level of probability of being exceeded. The
characteristic crack width represents the most probable
average crack width.
Keywords:- Reliability analysis, Flanged beam, Monte-
Carlo Simulation, Histogram, Probability distribution
curve.
I. Introduction
A. Characterization
Characterization is the process of establishing the
statistics of a random variable. This randomness is
characterized by different probability distribution function.
To generate the sample of random variable numerical
simulation methods are used and probability modeling is
done. After knowing the distribution which the random
variable follows representative values like characteristic
value which has the probability of exceedence can be
defined.
IS 456-2000 considers the characteristic values as
that value which has 5% of being exceeded or being
lowered. With the advent of fast digital computers
generation of random variables by numerical simulation
has become possible.
IS 456-2000 has not proposed characteristic crack
width. It is required to keep the probability of exceedence
as small as possible. The importance of the structure,
quality control in construction should be considered in
selecting the value of probability of exceedence.
The characteristic value of a random variable
following normal distribution is given by, +k*
Where, is the mean value and is the standard
deviation. The factor k is suitably selected to represent the
probability of exceedence or lowering.
 National Conference on Computer Aided Analysis and Design of Structures (CAADS-2016) Paper ID: CAD34

B. General Theory of Serviceability (No visible cracking)

The structure to be designed on the basis of limit state Pre-stressed: Type 3 0.1 to 0.2
should satisfy the following limit states: mm
Reinforced concrete normal design 0.3
a) Limit state of collapse (safety requirement): mm
The following limit states of collapse are
considered in design
i. Limit state of collapse in flexure Table 1: Classification of Structures According To
ii. Limit state of collapse in Crack Width (IS 1343-1980)
compression
iii. Limit state of collapse in shear II. Reliability analysis
iv. Limit state of collapse in torsion
b) Limit state of serviceability:
A. General
The limit states of serviceability consists of
i. Excessive deflection
Reliability is defined as the probability of given item
ii. Premature or excessive cracking
performing its intended function for a given period of time
iii. Corrosion
which encounters certain operating conditions. Reliability
iv. Excessive Vibration
goal can be expressed as, Resistance Action.
Out of these, the important limit states of
serviceability are excessive deflection and cracking. It is
Here resistance is nothing but the load carrying
well known that in modern practice, structural failures are
capacity and the action may be the end results that include
very common in terms of serviceability and are relatively
deflection, cracking, moments etc.,
rare in terms of safety. In particular it is the serviceability
limit state of durability that calls for particular attention, The main aim of structural reliability analysis is to
indeed all over the world. The problem of inadequate quantify the reliability of structures under the consideration
durability is linked not only to such factors as improper of the uncertainties associated with the resistance and
making of concrete, attack from the environment and loads. The uncertainties that must be considered are the
corrosion of reinforcement, but also to inadequate cover to physical uncertainties, statistical uncertainties and the
the reinforcement, improper detailing and inadequate sizes model uncertainties. The physical uncertainties include the
for structural members, resulting in excessive deflections loading environment, the material properties and the
and crack widths. geometry of the structures. Statistical uncertainties may due
to incomplete statistical information and the model
C. Serviceability Limit States: Crack-Width
uncertainties are due to the idealized mathematical
description used to approximate the actual physical
behavior of the structure.
When we talk about structures, it is always the
aesthetics and durability that matters the most. These things
B. Monte Carlo Simulation Technique
get influenced by the amount of cracks that are developed
in the structure. One can always eliminate such cracks by Without doing physical, real and actual experiment simply
using requisite number of small diameter bars which are by knowing its probability distribution a set of random
well distributed in order to create ample amount of flexural variable values can be created which is referred as
reinforcing steel which controls the phenomenon of simulation. A sample from a Monte Carlo simulation is
cracking. The number of bars can always be determined similar to a sample of experimental observation; in the
depending upon the minimum spacing requirements (SMIN). form of histograms these results of the simulation process
Thus, by creating an area of flexural steel by providing can be presented. Main objective of Monte Carlo
required number of bars will effectively control crack simulation is the generation of random numbers from
widths. prescribed probability distributions, for a given set of
generated random numbers, the simulation process is
The allowable crack width in concrete depends on the
deterministic. To generate normal variables Box Muller
environmental condition to which it is exposed.
technique is used.
Allowable
Type of concrete structure u1= 2ln (1/v1)1/2cos( 2v1 )
crack width
u2= 2ln (1/v2)1/2 sin( 2v2 )
Pre-stressed: Type 1
0 y1 = + u1
(No tension in the section)
y2 = + u2
0 to 0.1
Pre-stressed: Type 2
mm

Department of Civil Engineering, SDMCET, Dharwad-580002

 National Conference on Computer Aided Analysis and Design of Structures (CAADS-2016) Paper ID: CAD34

Where, a = Distance from compression edge to the point where

crack width is calculated.acr = Distance from the point
1 and 2 are the set of values for a set of random considered to the nearest longitudinal bar.
numbers 1 and 2. and are the mean and standard
deviation of corresponding variables under consideration. Surface crack width for the above considered
flange beam is calculated at 4 different locations.
III. Mathematical formulation Point A Directly under the rebar
The deterministic formulation is needed for the
Point B At the bottom corner
characterization of randomness in crack width in RC
beams. The following articles describe the mathematical Point C Mid way between two rebar
formulation as per the procedure suggested in IS 456-2000
Point D At 2/3rd (d-x) from the neutral axis
A. Introduction to T Beams
Table 1: acr for various points under consideration
T Beams are more commonly used in practice
than the rectangular sections since the slab above the beam
participates with the structural behavior of the beam. The acr Points
slab which is in the compression side of the T beam is
stressed laterally in compression. They resist the acr = Cmin A
compression forces. The tension forces are taken up by the acr = (Cmin2 + Cmin2)1/2 B
tension steel provided in the bottom web of the beam.
spacing 2 C
acr = ( ( ) + Cmin2 )1/2
2

1 2 D
acr = (( (d x) ) + Cmin2 )1/2
3 2

C. Basic design variables

Table 2: Statistics of basic design variables

Figure 1. Cross-section of a T Beam Sl no Design Mean value in mm Cov % Type of

variable distribution
Only between the supports, the T beam acts as a 1 D 300 650 5 Normal
flanged beam where positive bending moment prevails and 2 d 250 - 600 5 Normal
the slab is under compression whereas over the supports, 3 bf 1813.33 - 2146.67 3 Normal
negative bending moment prevails and the slab is under 4 bw 230 3 Normal
tension and hence the beam acts as a rectangular beam. 5 df 125 5 Normal
6 fy 415 10 Normal
7 fck 20 15 Normal
8 cl 50 10 Normal
B. Procedure for the Calculation of Surface
Crack Width of Singly Reinforced T Beam 9 Ast As per design 5 Normal
10 Asc As per design 5 Normal
Maximum bending moment under service load is computed
and the area of steel is calculated. Depth of neutral axis for
cracked section is determined and then the moment of Case 1:Neutral axis lying within flange (x <Df)
inertia of cracked section is calculated. By finding strain
For analyzing characterization of randomness in
due to applied load at the level of steel and at the level of
limit state of cracking for a RC T beam, a beam of effective
soffit1, Surface crack width for a section under applied
flange width bf =1630mm, breadth of web = 200mm, depth
load as per IS 456-2000 is computed.
of flange Df = 100mm, total depth D = 350mm, having a
3 span of 5m with effective cover of 30mm is considered.
Wcr =
1 +
2 ()
()
Compression strength of concrete fck = 20N/mm2, Yield
strength of steel fy = 415N/mm2 is used.
Where,
(B(Dx)(ax)
m = 1 -
3EsAst(dx)

Department of Civil Engineering, SDMCET, Dharwad-580002

 National Conference on Computer Aided Analysis and Design of Structures (CAADS-2016) Paper ID: CAD34

Table 3: Characteristic crack width at point A D. Histogram and Probability distribution

diagrams for surface crack width of singly
SL Service Mean Standard COV Characteristic reinforced beam
no load deviation in % crack width
KN/m Wch =+1.645
1 26 0.108 0.0101 9.34 0.125 mm
2 28 0.121 0.0109 8.43 0.1376 mm
3 30 0.134 0.0105 7.884 0.1508 mm
4 32 0.146 0.0111 7.62 0.164 mm

Table 4: Characteristic crack width at point B

SL Service Mean Standard COV Characteristic

no load deviation in % crack width
KN/m Wch =+1.645
1 26 0.144 0.0142 9.85 0.167 mm
2 28 0.162 0.0146 9.04 0.186 mm
3 30 0.177 0.0152 8.602 0.202 mm Figure 2: Histogram and probability
4 32 0.194 0.016 8.21 0.221 mm distribution curve for load 26kN/m under
point A
Table 5: Characteristic crack width at point C

SL Service Mean Standard COV Characteristic

no load deviation in % crack width
KN/m Wch =+1.645
1 26 0.177 0.0302 17.04 0.2272 mm
2 28 0.198 0.033 16.78 0.252 mm
3 30 0.219 0.036 16.70 0.280 mm
4 32 0.239 0.0373 15.64 0.3004 mm

Table 6: Characteristic crack width at point D

SL Service Mean Standard COV Characteristic

no load deviation in % crack width
KN/m Wch =+1.645 Figure 3: Histogram and probability
distribution curve for load 26kN/m under
1 26 0.158 0.027 17.58 0.204 mm point B
2 28 0.177 0.030 17.10 0.227 mm
3 30 0.194 0.032 16.51 0.246 mm
4 32 0.212 0.034 16.17 0.269 mm

Figure 4: Histogram and probability

distribution curve for load 26kN/m under
point C

Department of Civil Engineering, SDMCET, Dharwad-580002

 National Conference on Computer Aided Analysis and Design of Structures (CAADS-2016) Paper ID: CAD34

Figure 5: Histogram and probability

distribution curve for load 26kN/m under point D

E. Matlab

It is a high-performance language for technical

computing. It integrates computation, visualization, and
programming in an easy-to-use environment where
problems and solutions are expressed in familiar
mathematical notations.

Matlab program has been developed for singly

reinforced and doubly reinforced beams under different
locations of neutral axis for the determination of surface
crack width. K-S test is used to check the goodness- of- fit.

F. Conclusions

The randomness in surface crack width is

characterized by the normal distribution and the
characteristic crack width is derived as that value which has
a probability of exceedence of 5%, for RC T beams. This
will help in developing the durability based design concept
of RC members.

References

[1] R. Ranganathan (1990) Reliabilty Analysis and

design of structures Tata McGraw-Hill
Publication Company Limited, New Delhi.
[2] IS 456-2000, Indian Standard code of practice for
Plain and Reinforced Concrete
[3] SP 16:2000, Design Aids for reinforced concrete,
Bureau of Indian Standard, New Delhi.
[4] Dr. Punmia, Ashok Kumar Jain and Arun Kumar
Jain,Limit State Design of RCC .

Department of Civil Engineering, SDMCET, Dharwad-580002